mixed layer height measurements from doppler lidar using a ...mixed layer height measurements from...
TRANSCRIPT
Mixed layer height measurements from Doppler lidar using a composite method
T. A. Bonin1,2, B. J. Carroll3, R. M. Hardesty1,2, W. A. Brewer2
1Cooperative Institute for Research in Environmental Sciences2National Oceanic and Atmospheric Administration, Chemical Sciences Division
3Joint Center for Earth Systems Technology/Univ. of Maryland Baltimore County
INFLUX Doppler lidar deployment
• Halo Streamline lidar deployed on roof of Ivy Tech Building (4 stories above ground) in Aug 2013 until June 2015
• Lidar was upgraded at the end of 2015, and Halo Streamline XR was redeployed at same location in January 2016 to present
• Motivation: Measuring greenhouse gas emissions from city• Need wind profile and MH
Halo
HRDL
Motivation for composite fuzzy-logic technique
• Backscatter alone is not always sufficient to determine mixing height, especially when residual layer is present
• Variance alone may lead to a high determination of the mixing height, especially when non-turbulent wavelike motions are present
• Need to use information from multiple scans to overcome minimum range issues
Scanning strategy for INFLUX
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
Use relation between HF w’2 (T<1 min, f>0.017 Hz) to total w’2 to differentiate turbulent and non-turbulent motions
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify proxies for mixing if it is near zi_fg• Peaks in wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
σw2 [m2 s-2] σu
2 + σv2 [m2 s-2] 10° VAD σvr
2 [m2 s-2]
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify proxies for mixing if it is near zi_fg• Peaks in wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
σw2
σw2 [m2 s-2]
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify proxies for mixing if it is near zi_fg• Peaks in wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify other indicators of mixing near zi_fg• Wind shear • Large variance or gradients in RCI
5) Determine final estimate for top of mixed layer
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify other indicators of mixing near zi_fg• Wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify other indicators of mixing near zi_fg• Wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer & uncertainty
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
MLH Detection Overview
1) Detect gravity waves and other non-turbulent sub-mesomotions
2) Combine data from all useful scans using fuzzy logic:• σvr
2 from each VAD scan • σu2, σv
2 from RHI scans• σu
2, σv2 from shallow stares • σw
2 from vertical stares
3) Identify a first guess for the top of the mixed layer (zi_fg) based on fuzzy aggregation
4) Fuzzify other indicators of mixing near zi_fg• Wind shear • Large variance or gradients in SNR
5) Determine final estimate for top of mixed layer & uncertainty
6) Flag the final estimate:• Is it raining? • Can we see the top of the ML?• Is the ML cloud-topped? • Is ML below minimum height?
Verification of MLH with aircraft observations5/13/16 in Indianapolis
AscentDescent
We thank Paul Shepson, Olivia Salmon, and the entire Atmospheric Chemistry group Purdue University for taking profiles over the lidar site and providing the independent observations for comparison
Annual Variation of MLH
Normalized Diurnal variability in MLH in Indianapolis
Morning transition Evening transition
MLH evolution depends on mean wind speed
Morning transition Evening transition
Summary
• A composite fuzzy logic algorithm has been developed and applied to different Doppler lidar systems to continuously detect the MLH at high temporal resolution (15-20 min)• Uses inputs of velocity variances (turbulence), backscatter intensity, and wind
profiles from all scans• Gravity waves and other non-turbulent motions are identified and flagged for
exclusion from analysis
• We have applied this algorithm to other Doppler lidars at in different locations (Oregon, California, Las Vegas, Alaska)• Algorithm adjusts to use whatever data it can get; do not need the scanning
pattern discussed here
• Ongoing efforts to validate MHs through intercomparison with other instruments and NWP output